This article is about a roadmap for reshaping data governance.

The previous article of this series discussed why organizations may need to reassess their existing frameworks and outlined what should be done. However, the most critical question remains: how to implement these changes effectively. This article provides a high-level approach for developing a new governance framework or adjusting an existing one.

Introduction to the Roadmap Approach

My pragmatic approach to implementing a governance framework for data management is shaped by my career journey in the field. I initially built a governance framework from the ground up without prior knowledge of the subject. I relied on online research to learn and apply different aspects in practice. At the time, I was unaware of DAMA-DMBOK publications and only discovered them six years later—by which point my framework was already in place. Comparing the theoretical insights from DAMA-DMBOK with my hands-on experience revealed significant shortcomings from an implementation perspective. To address these gaps, I developed the O.R.A.N.G.E. Data Management Framework (illustrated in Figure 1) and have detailed its application in several books. While these books cover different topics and bear distinct titles, they all build upon the evolving O.R.A.N.G.E. DMF and demonstrate its wide range of applications—from establishing governance frameworks for data and metadata management to integrating them with governance frameworks for AI management.

Figure 1: The O.R.A.N.G.E. Data Management Framework.

Figure 1: The O.R.A.N.G.E. Data Management Framework.

The O.R.A.N.G.E. Data Management Framework (DMF) diagram in Figure 1 illustrates only its core components. However, understanding how these components interact and contribute to the full lifecycle of a governance framework is crucial.

Figure 2 depicts how the foundational blocks of the O.R.A.N.G.E. DMF evolve into a comprehensive governance framework lifecycle and roadmap. This transformation supports the architecture for (meta)data, AI, and related risk management, ensuring a structured and integrated approach and an implementation guideline.

Figure 2: Figure 2: The O.R.A.N.G.E. DMF lifecycle and roadmap.

Figure 2: The O.R.A.N.G.E. DMF lifecycle and roadmap.

Key Phases in Establishing or Refining a Governance Framework for (Meta)Data and AI Management

This section briefly overviews the key phases involved in developing or refining a governance framework for (meta)data and AI management, along with the interdependencies between these phases.

ORGANIZE

The first phase, ORGANIZE, focuses on analyzing an organization’s business needs and the resources required for a data-related initiative. This phase ensures that the governance framework’s scope and implementation remain feasible. Several factors influence the scope, including the target data architecture, the breadth of data and metadata covered, the need to align (meta)data and AI practices, and compliance with data and AI regulatory requirements.

The key deliverables of this phase include long-term strategic plans, such as a governance strategy, strategic progress KPIs, a high-level framework design, the implementation approach and methodology, and a high-level roadmap. The chosen implementation method determines whether the organization will design and implement the framework concurrently or sequentially.

While defining the initiative’s scope, an organization must incorporate elements from other phases. For instance, it should assess the current maturity level and establish target maturity goals for existing capabilities using techniques from the GROW phase. Conducting a gap analysis helps in shaping long-term plans.

Additionally, organizations should establish strategic KPIs to measure the progress and success of framework implementation. Techniques from the NAVIGATE phase support this effort.

Once the framework’s scope and long-term plans are defined, the organization can proceed with designing the framework architecture (RENDER phase) and implementing it (ACTIVATE phase).

A common approach is to first design specific capabilities and then begin implementation with some overlap between the two processes. The critical aspect is selecting the necessary capabilities and developing an integrated implementation plan. The following article in this series will explore this topic in greater detail.

RENDER

The primary objective of this phase is to develop a comprehensive governance framework that encompasses the following key components, as shown in Figure 3.

Figure 3: Key Areas of a Governance Framework.

Figure 3: Key Areas of a Governance Framework.

Enterprise-Wide Governance Framework

Let’s start by examining the essential deliverables.

Operating Model

The design of the operating model depends on several critical factors, including the organization’s business model, organizational structure, size, geographical distribution, target data, application, and technology architectures. Organizations can adopt centralized, decentralized, or hybrid governance models, with the key distinction being how tasks and responsibilities are distributed between central and local data management offices.

Organizational Structure

A well-defined governance framework requires an organizational structure that includes both governing bodies and data-related roles.

  • Governing Bodies

Every organization operates multiple business governing bodies. Even within data and IT management, examples include the Enterprise Architecture Board, IT Change Board, and Data Governance Board. Optimizing the structure of these boards is crucial, as the same business leaders often participate in multiple governing bodies.

  • Data-Related Role Architecture

In the data and AI landscape, various roles are necessary. These roles fall into two main categories:

  • Functional roles (part of the formal organizational structure), such as the Chief Data Officer (CDO).
  • Virtual roles (assigned to functional roles), such as data owners and data users who fulfill specific responsibilities

Additionally, it is essential to differentiate roles based on their professional background, distinguishing between business, data, and IT functions.

Capability-Specific Governance Framework

Each lower-level data management capability requires its governance framework to function effectively. This includes several key components:

  • A defined list of artifacts and necessary inputs
  • Policies and standards
  • Processes linked to roles with clear accountability levels (RACI or RASCI models)
  • Requirements for supporting IT tools
  • Other assets

For example, consider the data quality capability:

  • The primary deliverable is ensuring data meets the required quality standards. Several enabling deliverables support this goal, including data quality (DQ) dimensions, DQ requirements, DQ checks and controls, and DQ validation.
  • Mandatory inputs include data flows and lineage, metadata management, data models, and business glossaries.
  • Policies and standards define how DQ will be implemented.
  • Processes linked to roles (using RACI) ensure the effective execution of policies and the delivery of necessary artifacts.
  • Modern IT tools can automate many DQ processes, but defining clear requirements is essential for effective and sustainable implementation.

Collaboration Mechanism Between Data, Metadata, and AI Governance Capabilities

The governance framework must establish an effective collaboration mechanism across various capabilities, including data management, metadata management, AI governance, and data and AI risk management. The configuration of this framework will depend on the level of integration between these capabilities.

When designing governance frameworks, methodologies from the NAVIGATE phase (performance management) and GROW phase (maturity measurement) should be leveraged to define the required scope in detail.

ACTIVATE

While defining the scope and long-term plans, the organization should have established the method and approach for implementing the governance framework architecture. Several factors influence this decision:

  • Structural level: Integration or separation of (meta)data, AI, and risk frameworks
    The organization must determine whether to consolidate these frameworks into a unified structure or maintain separate systems tailored to the specific needs of data and AI governance.
  • Execution level: Simultaneous or independent development
    The timing of implementation must be carefully planned. The organization can choose between a concurrent rollout, where multiple components are implemented simultaneously, or a phased approach, where elements are introduced sequentially based on capacity and priorities. Both approaches can be applied to either integrated or separate frameworks.
  • Alignment level: Relationship between design and implementation phases
    This factor ensures that the conceptual design of the governance framework is effectively translated into practice, aligning strategic objectives with real-world execution. A well-structured alignment between design and implementation helps prevent gaps between vision and operational reality.

NAVIGATE

This phase establishes a performance management capability and a governance control framework. The design of the performance management capability depends on several key factors:

  • Phase in establishing a governance framework: During the scoping, design, and implementation stages, the primary focus is developing Progress KPIs that track implementation milestones. Once enterprise-wide and capability-specific frameworks are in place, the focus shifts to Performance KPIs, which assess the effectiveness and efficiency of the governance framework in operation.
  • Organizational level: The number of organizational levels—strategic, tactical, and operational—determines the types of KPIs required and the level of detail needed at each level.
  • Direction of setting and linking different KPIs
  • Both top-down and bottom-up approaches can be applied when defining and aligning KPIs, ensuring that performance measurement is integrated across all levels of governance.

GROW

The NAVIGATE and GROW phases are closely connected. The relationship between them is straightforward: maturity measurement helps assess the current state of a framework’s architecture and define its target state. Gap analysis informs the development of long-term plans during the SCOPE phase and medium- and short-term plans during the RENDER and ACTIVATE phases. The outcomes of these plans and actions are reflected in the KPIs.

The O.R.A.N.G.E. DMF provides three methods for measuring maturity:

  1. Perform a free maturity scan

A free online maturity scan is available at:
//datacrossroads.nl/free-resources/#maturityscan

  1. Conduct a detailed data management maturity assessment

Organizations can undergo a maturity assessment using the O.R.A.N.G.E. (Meta)Data Management Maturity Model, which evaluates ten core (meta)data management capabilities and provides a tailored execution plan. The assessment is available at:
//datacrossroads.org/survey/init. Those interested in this assessment can contact me directly.

  1. Develop a custom maturity model

A maturity model can be developed based on an organization’s specific requirements.

EVOLVE

This phase establishes the foundational principles for creating a flexible, scalable, and adaptive governance framework architecture. That is why, in Figure 2, it serves as the foundation for all other phases.

DISTINGUISHING FEATURES OF THE O.R.A.N.G.E. DMF

Based on the aspects discussed above, the O.R.A.N.G.E. DMF has several distinguishing characteristics:

  • It views data management as a set of interconnected capabilities, where the artifacts of one capability serve as inputs for others.
  • This interlinked capability approach facilitates the development of an integrated implementation plan, aligning governance framework development with IT tool implementation.
  • The framework consists of multiple building blocks:
    • A model of a data management capability and its core components
    • A method for integrated implementation of a governance framework across core data management capabilities
    • Models and methodologies for assessing maturity and performance of the governance framework
    • A model for organizing core data management artifacts into a data management knowledge graph
  • The framework, particularly its implementation component, can be applied to industry-specific or custom-developed governance frameworks used by different organizations.
  • It allows for data management and AI governance capabilities customization based on an organization’s unique needs and resources.
  • It is a ready-to-use methodology that includes multiple templates.

The third article in this series will provide a detailed approach for designing and implementing the governance framework architecture for (meta)data, data, AI, and related risk management.